Ambient temperature and preterm birth risk in New York City: A time-series analysis

Abstract

BACKGROUND AND AIM Recent studies suggest an association between daily variation in ambient temperature during pregnancy and preterm birth (PTB). Our objective is to examine the short-term effects of heat and cold on PTB risk from 2008 to 2016 in New York City (NYC). METHODS We applied quasi-Poisson regression with distributed lag nonlinear models to estimate the cumulative relative risk (CRR) of daily counts of spontaneous PTB (<37 weeks gestation) and outdoor temperature, and to estimate the fraction of PTBs attributable to temperature, by season, using NYC Vital Statistics birth records and National Weather Service daily temperature data at LaGuardia airport. Models were adjusted for within-season trends and daily number of pregnant people at risk for preterm delivery. Heat and cold were defined as the 95th versus 50th percentile of maximum temperature during the warm season and 5th versus 50th percentile of minimum temperature during the cold season, respectively. Maternal race/ethnicity was assessed for effect modification. RESULTS Our analysis consisted of 38,593 spontaneous PTBs. Heat was not associated with PTB during the warm season [CRR: 0.98 (95% CI: 0.92, 1.05)]. There was a nonlinear association between cold and PTB. During the cold season, the CRR was 1.08 (0.99, 1.17) for cold up to three days before delivery (lag 0-3). The estimated fraction of PTBs attributable to cold was 9.3% (-0.6%, 17.4%), which corresponded to 81 (-2, 154) cold-attributable PTBs over the 9-year study period. Among non-Hispanic Blacks, the CRR was 1.20 (1.04, 1.39) for cold over lag 0-3 days, and an estimated 23.2% (10.3%, 34.3%) of non-Hispanic Black PTBs, equivalent to 69 (29, 105) PTBs, were attributable to cold. CONCLUSIONS Heat had no measurable impact. Cold potentially increased risk of PTB. Our finding for the impact of cold on non-Hispanic Black PTBs warrants additional research. KEYWORDS Preterm birth Temperature